Heat exposure and productivity loss among construction workers: a meta-analysis

Abstract

Background

Global warming is raising increasing concerns about its impact on worker productivity, particularly in industries like construction where outdoor physically demanding jobs are prevalent. This study aimed to perform a meta-analysis to assess the existing evidence on the impact of heat exposure on productivity loss among construction workers.

Methods

We conducted a comprehensive literature search across six databases—Web of Science, PubMed, Embase, Scopus, ScienceDirect, and IEEE—covering the period from database inception to September 18, 2024. The Joanna Briggs Institute (JBI) critical appraisal checklist was used for quality assessment. A random-effect model meta-analysis was performed, and publication bias was evaluated by Egger’s and Begg’s tests.

Results

From an initial pool of 1209 studies, 14 met the inclusion criteria, representing data from 2387 workers. Our findings indicate that 60% (95% CI: 0.48–0.72, p < 0.01) of construction workers exposed to elevated temperatures experienced significant productivity loss. Productivity loss was more pronounced when the Wet Bulb Globe Temperature (WBGT) exceeded 28 °C or when ambient temperatures surpassed 35 °C. Furthermore, workers aged over 38 (proportion = 0.61, 95% CI: 0.49–0.72) and teams with female workers (ratio = 0.74, 95% CI: 0.60–0.87) were more susceptible to productivity loss.

Conclusions

This review highlights heat exposure as a significant factor affecting productivity in the construction industry. We recommend prioritizing the protection of vulnerable groups such as women and older workers, developing innovative technologies and equipment for working in hot conditions, and improving the working environment to safeguard workers’ health and productivity. Further research is needed to investigate the long-term health impacts of heat exposure and develop strategies for optimizing microclimate management in construction settings.

Introduction

In the context of global climate change, the increasing frequency and intensity of extreme heat events pose a severe threat to outdoor workers, particularly in labor-intensive sectors like construction. This industry requires physically demanding work, often performed in unprotected environments. Construction workers engaged in tasks such as scaffolding, reinforcement fixing, structural steel erection, and concrete pouring under direct sunlight are frequently exposed to high temperatures for extended periods, which disrupts the production process and results in decreased work productivity [1].

Previous studies have shown that heat exposure reduced the productivity of construction workers. For instance, Li [2] conducted a study in Beijing during the summer of 2014, measuring WBGT and labor productivity in two construction projects. The findings showed a 1℃ increase in WBGT led to a 0.57% decline in productivity. The impact varied depending on the workers’ age, experience, and skill level. Similarly, Sett [3], in her research on female brick factory workers in India, showed that productivity decreased linearly when maximum temperatures exceeded 34.98 °C. For each degree of increase, productivity losses amounted to approximately 2%. Lin [4], using public databases, analyzed historical meteorological, demographic, and economic data, concluding that construction workers were among the most vulnerable groups, with heat exposure leading to substantial productivity losses.

Productivity is a critical factor for the global construction industry, directly influencing economic growth, cost-effectiveness, project timelines, innovation, and environmental sustainability [5]. However, rising temperatures present a growing challenge to construction productivity. Heat exposure affects workers’ physical health through various physiological and psychological mechanisms, including dysregulation of body temperature, dehydration, electrolyte imbalances, cardiovascular strain, cognitive impairment, psychological stress, accelerated physical fatigue, and impaired immune function [6, 7]. These factors contribute to a significant reduction in productivity under extreme heat conditions.

Although many studies have investigated the impact of heat exposure on construction workers’ productivity, there is considerable variation in sample sizes and geographic regions across these studies. This variability complicates the task of forming a comprehensive understanding of the issue. Therefore, we conducted a meta-analysis to systematically integrate the existing literature, synthesizing data from multiple regions and using quantitative methods to assess the prevalence of productivity loss due to heat exposure. We also examined how individual characteristics, such as age and gender, influenced the results. Additionally, this review highlights limitations in the current body of evidence and offers directions for future research.

Methods

This review adhered to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) checklist. The protocol was registered at the Prospective Register of Systematic Reviews (PROSPERO) under CRD 42,023,449,885.

Search strategy

We employed a comprehensive search approach across six reputable electronic databases: Web of Science, PubMed, Embase, Scopus, ScienceDirect, and IEEE, which are recognized sources of high-quality publications in construction engineering and management. The search utilized the following keywords: “(heat exposure OR heat stress OR heat wave) AND (productivity OR work time OR task analysis) AND (Construction worker* OR Construction labo* OR construction Industry)” for relevance. We limited our search to English-language papers published up to September 18, 2024, yielding a collection of peer-reviewed scientific articles across various fields.

Inclusion and exclusion criteria

Inclusion criteria

1. Participants must be workers in the construction industry, with no specific limits on gender or age. This information will be recorded for subsequent stratified analysis.

2. Studies must include indicators of heat exposure, including ambient temperature or WBGT. While there was no limit on the level of heat exposure, studies must indicate that they were conducted during hot seasons or in high temperatures.

3. Studies must report productivity indicators, including work output, percentage of work time lost, and self-reported productivity losses by the workers.

4. Full texts of the literature must be available in English, or reliable English abstracts and critical data can be extracted.

Exclusion criteria

1. Review articles, editorials, letters, conference abstracts, case reports and expert opinions.

2. Research not involving individuals working in the construction industry.

3. Studies that do not report on productivity or related indicators.

4. For articles published multiple times in the same dataset, only the most complete or latest version of the article was selected.

Literature screening and data extraction

We conducted a comprehensive search across the six electronic databases, applying pre-set keyword combinations and search strategies. The results were managed using EndNote to prevent duplicate inclusions. Two independent reviewers conducted a preliminary screening of all the retrieved literature titles and abstracts, respectively. Each reviewer evaluated the eligibility of documents based on the preset inclusion criteria. In cases of disagreement, a third independent arbiter intervened to facilitate consensus. Full texts of the literature that passed preliminary screening were obtained, and if unavailable, authors or publishers were contacted for assistance. Two independent reviewers thoroughly evaluated the full texts to confirm adherence to the inclusion criteria, focusing on research quality and data completeness. Data was systematically organized using Excel for the final studies.

Data extracted from all studies included author names, year of publication, number of participants, age, gender, temperature, and outcomes. Heat exposure was defined as the exposure of individuals or populations to high temperatures, high humidity, or other conditions that exacerbate heat stress. Data on WBGT or ambient temperature were extracted. Lost productivity was defined as lost labor time, decreased performance, and absences due to heat exposure. For each study, the prevalence of productivity loss due to heat exposure was calculated as the number of workers who experienced heat-induced productivity loss, either through monitoring or self-reports, relative to the total sample size. Additional information was requested from the authors by email when necessary.

Risk of bias

The methodological quality of the selected studies was assessed using the Statistical Assessment and Review Tool (MAStARI) from the Joanna Briggs Institute, specifically designed for cross-sectional studies [8]. Two independent reviewers evaluated each study separately and scored it item by item according to JBI-MAStARI criteria. A score of “Yes” received one point, while “No”, “Not specified”, and “Inappropriate” were scored zero. The risk of bias (ROS) was categorized based on scores: High ROS when the study scored less than four, moderate ROS when it reached a score of four but less than six, and low ROS when the score reached up to six. Reviewers compared their assessments and addressed discrepancies through discussion. If consensus was not achieved, a third independent arbiter made the final decision.

Statistical analysis

Statistical analyses were conducted using R 4.4.1. We calculated event rates (P) and standard errors (SE) using the formulas and . The construction of the confidence interval was based on the normal distribution assumption. To address issues with small sample sizes and extreme ratios (close to 0 or 1), we applied a double arcsine transformation, enhancing normality and stabilizing variance for comparative analysis.

Heterogeneity among studies was assessed using the Q test and I² values. If the P-value of the Q test is greater than 0.10, there is no significant heterogeneity. A Q test P-value > 0.10 indicated no significant heterogeneity, while a P-value < 0.10 suggested significant heterogeneity. An I² value < 50% indicated low to moderate heterogeneity, and ≥ 50% indicated significant heterogeneity. We employed a fixed-effect model for low heterogeneity (P > 0.10 and I² < 50%) and a random effects model for high heterogeneity (P ≤ 0.10 and I² ≥ 50%).

Subgroup analyses were performed to explore sources of heterogeneity, focusing on clinical (participant characteristics and outcomes), methodological (study design and ROB), and statistical heterogeneity. The 14 articles were categorized into six groups based on age, gender, country, study duration, heat exposure, and productivity measurement methods. Methodological heterogeneity was reduced by grouping studies into three categories: group 1 (low ROB convenience samples), group 2 (moderate ROB convenience samples), and group 3 (randomly recruited participants).

Sensitivity analyses assessed the stability of findings by sequentially excluding studies from the meta-analysis. The impact of each study on the total effect size was evaluated, and studies missing age and gender information were excluded in subgroup analyses. Cook’s Distance statistics were calculated to identify influential studies, with a value > 1 indicating significant impact. Finally, potential publication bias was assessed using funnel plots, along with Egger’s and Begg’s tests. A P-value < 0.05 was considered indicative of significant statistical divergence.

Results

Literature search

Our database search yielded 1,209 articles, and after a rigorous screening process, we narrowed down the selection to 14 papers (Fig. 1).

Fig. 1.

Literature selection process [9]

Basic characteristics of the included studies

Table 1 presents key characteristics of the 14 articles (involving 2387 participants) included in this review. These studies covered both high-income and middle-low-income country countries, with the majority coming from India [3, 10–12], China [2, 13–15], Thailand [16], Australia [17], Italy [18], Iran [19], Saudi Arabia [20] and Sweden [21].

Table 1.

Basic information about the studies included in the review

Inclusion of literature	Year of publication	Country	Sample size	Case	Heat exposure (average)
Langkulsen et al.	2010	Thailand	5	3	WBGT 27.1℃
Sett et al.	2014	India	32	107	WBGT 24.23℃
HAJIZADEH et al.	2015	Iran	184	126	WBGT 31.84℃
Venugopal et al.	2015	India	52	44	WBGT 28.7℃
Li et al.	2016	China	16	2	WBGT 27.19℃
Chinnadurai et al.	2016	India	4	3	WBGT 30.3℃
Yi et al.	2017	China	14	2	WBGT 29.39℃
Lundgren et al.	2018	India	87	42	WBGT 27℃
Moda et al.	2018	Saudi Arabia	100	50	Ambient temperature 30℃
Zander et al.	2018	Australia	114	60	Ambient temperature 30℃-
Messeri et al.	2019	Italy	78	61	WBGT average 28℃-
Larsson et al.	2021	Sweden	232	139	Ambient temperature 25℃
Han et al.	2021	China	318	202	Ambient temperature 28℃
Fang et al.	2021	China	1063	650	Ambient temperature 38.83℃

For temperature measurement, nine studies used WBGT [2, 3, 10–13, 16, 18, 19], an indicator that combines the effects of environmental temperature, relative humidity, wind speed, and solar radiation. WBGT is widely accepted by institutions like the American Conference of Government Industrial Hygienists, the National Institute of Occupational Safety and Health, and the International Organization for Standardization, as a key indicator for setting heat safety limits in industrial workplaces [22, 23]. The remaining five studies used ambient temperature [14, 15, 17, 20, 21]. In the studies that reported WBGT, the average WBGT exceeded 24 °C, with the highest reaching 31.84 °C. In the studies that reported WBGT, the average ROS exceeded 24 °C, with the highest reaching 31.84 °C. All the studies were conducted during high-temperature seasons, with construction workers exposed to significant heat during their tasks. To further understand how different levels of heat exposure affected productivity loss, we conducted a stratified temperature analysis, as detailed in the results section.

Productivity refers to the relationship between output and related input in the production process. Construction labor productivity is widely regarded as the ratio of output to input, and the American Association of Cost Engineers (AACE International ) [24] also defines construction labor productivity as “the output rate per unit time or workload”. This study considered the effects of heat exposure on construction workers, focusing on outputs, including work completion time, reduced productivity rates, and absenteeism due to heat exposure. Loss of labor time, performance, and absenteeism can all lead to reduced productivity for construction workers. The studies varied in how they measured productivity loss.

Two studies collected qualitative data on heat exposure, potential health impacts, productivity loss, and coping mechanisms through the ‘high occupational temperature health and productivity suppression’ programme (Hothaps questionnaire) to assess the effects of high temperatures on worker productivity [10, 12]. The Hothaps questionnaire is designed to describe and quantify the impact of high temperatures on occupational health and work capacity in different regions of the world, taking into account future climate change [25]. One study showed that the decline in labor productivity was caused by failure to meet expected goals, absenteeism, or indirectly by physical factors like heat exposure [10]. Another study reported that productivity decreased, tasks took longer time to complete, and overtime was required to meet targets [12]. Two studies [2, 13] measured labor productivity using the method proposed by AACE International, categorizing work activities into three categories: [1] Direct task allocation work that requires specific efforts, or the use of tools/equipment that contribute productively to the completion of task; [2] Indirect task support work or assistance, which is not conducive to the completion of the task; [3] Non-productive - personal time and non-usage time caused by work stoppages for any reason, to collect and record productivity-related data. One study showed that direct working hours accounted for 74% of time, indirect hours for 15%, and non-productive hours for 11% [2]. Another study showed that workers’ non-productive time, including personal time and downtime, made up 11% of total time [13]. One study involved interviews with four workers in various construction jobs and provided a questionnaire about the tasks they completed during their shifts. In addition to this, each worker was directly observed daily during the measurement process to verify the worker’s reported workload estimates and productivity. As a result, the productivity of the three people decreased [11]. One study used the cognitive ability questionnaire to assess workers’ responsiveness and labor productivity, where more than half of the workers experienced productivity loss [15]. The remaining studies used self-made or adapted questionnaires to measure productivity losses. Overall, all studies concluded that heat exposure negatively affected construction workers’ work capacity and productivity. Moreover, the variety of measurement tools led to variations in findings, prompting a subgroup analysis, detailed in the next section.

Risk of bias

A total of 6 studies presented low ROB, while 8 presented moderate ROB, as shown in Table 2 below.

Table 2.

Scoring results of included literature

Study	Q1	Q2	Q3	Q4	Q5	Q6	Q7	Q8	Scores	Risk of bias
Langkulsen 2010	N	Y	Y	Y	N	N	Y	Y	5	Mod
Sett 2014	N	Y	Y	Y	Y	Y	Y	Y	7	Low
HAJIZADEH 2015	Y	Y	Y	Y	N	N	Y	Y	6	Low
Venugopal 2015	N	Y	Y	Y	N	N	Y	Y	5	Mod
Li 2016	N	Y	Y	Y	Y	Y	Y	Y	7	Low
Chinnadurai 2016	N	N	Y	Y	Y	N	Y	Y	5	Mod
Yi 2017	N	Y	Y	Y	Y	Y	Y	Y	7	Low
Lundgren 2018	N	N	Y	Y	Y	N	Y	Y	5	Mod
Moda 2018	Y	Y	N	N	Y	Y	Y	Y	6	Low
Zander 2018	N	Y	N	N	Y	Y	Y	Y	5	Mod
Messeri 2019	N	Y	N	N	Y	N	Y	Y	4	Mod
Larsson 2021	Y	N	N	N	Y	Y	Y	Y	5	Mod
Han 2021	Y	Y	N	N	N	N	Y	Y	4	Mod
Fang 2021	Y	Y	Y	Y	Y	N	Y	Y	7	Low

Q1. Were the criteria for inclusion in the sample clearly defined?

Q2. Were the study subjects and the setting described in detail?

Q3. Was the exposure measured in a valid and reliable way?

Q4. Were objective, standard criteria used for measurement of the condition?

Q5. Were confounding factors identified?

Q6. Were strategies to deal with confounding factors stated?

Q7. Were the outcomes measured in a valid and reliable way?

Q8. Was appropriate statistical analysis used?

Meta-analysis results

Combined effect size

A comprehensive analysis of 14 articles revealed significant heterogeneity (I² = 90%, p < 0.01), indicating considerable variation across studies. Due to this, a random effects model was applied for data synthesis. We used a weighted average approach, combining information from different studies. The results showed that 60% (CI: 0.48 to 0.72) of construction workers exposed to heat experienced productivity loss (see Fig. 2), underscoring the negative impact of heat exposure on productivity in the construction industry.

Fig. 2.

Meta-analysis diagram

Subgroup analyses

Clinical heterogeneity

Subgroup analysis showed that construction workers aged 38 and older had a productivity loss rate of 61%, compared to 58% for those under 38. The productivity loss was 45% in male-only groups and 74% in mixed-gender or all-female groups. Construction workers in middle-low-income countries had a 58% productivity loss rate, compared to 64% in high-income countries. The study period included in the review was from 2010 to 2021, and the results of meta-analysis varied depending on the year. Studies conducted before 2016 reported a 63% loss rate, compared to 57% for studies conducted after 2016.

We also observed a non-linear relationship between heat exposure and productivity loss. When the Wet Bulb Globe Temperature (WBGT) was ≥ 28 and ambient temperature > 35, the productivity loss was 62%, while in cooler conditions (WBGT < 28, ambient temperature < 35), the loss was 59%. Productivity loss based on direct measurement was 13%, while self-reported loss through questionnaires or interviews was 67%. More specifically, there was little difference in self-reported results using HOTHAPS, the Cognitive Ability Questionnaire, and other adapted questionnaires. Significant heterogeneity (I²>50%) was found in subgroups related to countries, study periods, and heat exposure levels, while age, gender, and measurement methods were identified as the primary sources of heterogeneity, as shown in Table 3.

Table 3.

Meta-analysis of the rate of productivity loss in different subgroups of construction workers

Subgroups	Number of studies	heterogeneity test		effect model	Meta-analysis
		P-value	I² Value (%)		Proportion	95%- CI
Age
<38	6	< 0.01	94%	random	0.58	(0.30;0.84)
≥ 38	5	< 0.01	79%	random	0.61	(0.49;0.72)
Other	3	0.15	48%	random	0.56	(0.45;0.67)
Only male
Yes	6	< 0.01	87%	random	0.45	(0.25;0.66)
No	5	< 0.01	93%	random	0.74	(0.60;0.87)
Other	3	0.15	48%	random	0.56	(0.45;0.67)
Country
middle-low-income country	11	< 0.01	91%	random	0.58	(0.42;0.74)
high-income country	3	< 0.01	86%	random	0.64	(0.48;0.78)
Study period
Prior to 2016	8	< 0.01	92%	random	0.63	(0.43;0.81)
After 2016	6	< 0.01	83%	random	0.57	(0.43;0.71)
Heat exposure
WBGT ≤ 28℃/Ambient temperature ≤ 35℃	9	< 0.01	92%	random	0.59	(0.44;0.73)
WBGT > 28℃/Ambient temperature > 35℃	5	< 0.01	87%	random	0.62	(0.37;0.84)
Productivity measurement method
Self-report	12	< 0.01	88%	random	0.67	(0.58;0.75)
actual measurement	2	0.88	0%	random	0.13	(0.03;0.29)

Methodological heterogeneity

We grouped studies based on ROB and performed meta-analyses on each group (see Fig. 3). Forest plots showed high heterogeneity across studies, prompting the use of a random effects model. Group 1 (low ROB) had an inconsistency (I²) of 96% (CI: 0.93 to 0.98), p < 0.01, while group 2 (moderate ROB) had I² = 82.6% (CI: 0.64 to 0.92), p < 0.01, and group 3 (random samples) had I² = 78% (CI: 0.40 to 0.92), p < 0.01. Despite the heterogeneity, groups 2 and 3 were methodologically rigorous, reducing selection bias and contributing to the overall effect size, enhancing the stability and reliability of conclusions (Table 4).

Fig. 3.

a Meta-analysis diagram of group1. b Meta-analysis diagram of group2. c Meta-analysis diagram of group3

Table 4.

Test for heterogeneity and Meta-analysis of three groups

Group	Test for heterogeneity					Meta-analysis
						Fixed effects		Random effects
	Q	Df	Significance level	I²	95% CI	proportion	confidence interval	proportion	confidence interval
Group1	81.49	3	< 0.0001	96.3%	93.2%; 98.0%	64	0.61;0.66	46	0.10; 0.84
Group2	29.35	5	< 0.0001	83%	64.1%; 91.9%	65	0.60; 0.70	69	0.54; 0.82
Group3	13.47	3	0.0037	77.7%	39.7%; 91.8%	61	0.58; 0.65	59	0.51; 0.68

Sensitivity analysis

Sequential exclusion of each of the 14 studies revealed that P-values remained below 0.05, indicating a consistent combined effect size of around 60%, regardless of which study was excluded. This suggests that the results are robust, with no single study driving the high heterogeneity (Fig. 4a). Additionally, excluding three studies with missing information did not significantly change the effect size, reinforcing the robustness of our findings (Fig. 4b). Cook’s Distance values in all studies were below 1, indicating that no individual study disproportionately influenced the overall.

Fig. 4.

A Sensitivity analysis. B Forest plot following the exclusion of three studies

results.

Publication bias

A funnel plot showed some asymmetry, with more dots concentrated on the right and fewer scattered on the left (Fig. 5a), suggesting potential publication bias. However, Egger’s test (t = -0.29, p = 0.6222) and Begg’s test (z = -0.49, p = 0.6222) provided no statistical evidence of significant publication bias (Fig. 5b).

Fig. 5.

A funnel plot. B Egger’s test regression plot

Discussion

This meta-analysis included 14 studies and aimed to quantify the productivity loss among construction workers due to heat exposure.

Key findings and mechanisms of productivity loss attributed to heat exposure

Our results showed that up to 60% of construction workers experienced productivity loss when the average WBGT was above 24.23. This result was higher than Flouris’s meta-analysis [26], suggesting a more significant impact of heat exposure on construction workers compared to other occupations. The mechanism underlying this productivity loss is multifaceted, involving both physiological and psychological factors. Physiologically, heat stress triggers sweating and dehydration, which lead to electrolyte imbalances, fatigue, dizziness, and decreased concentration [27]. Furthermore, heat imposes strain on the cardiovascular system, accelerating blood circulation to dissipate heat, which in turn reduces blood flow to muscles, impairing strength and endurance [28]. In addition, discomfort and stress in a high-temperature environment can also hurt the psychological state of workers, increasing anxiety and irritability, thus affecting decision-making ability and work accuracy [29]. These physiological and psychological factors work together to reduce the productivity of construction workers significantly.

The discrepancy with Flouris’s results likely arises from differences in the occupational groups included. Flouris’s study encompassed a wider range of occupations, such as indoor office workers and farmers, whose work conditions differ significantly from construction workers in terms of intensity, flexibility, and protective measures [30, 31]. Office workers primarily engage in mental tasks with minimal physical exertion, while farmers, although exposed to outdoor conditions, have more flexibility in adjusting work hours and can seek natural shade. In contrast, construction workers face direct exposure to heat and sunlight, often in environments lacking proper ventilation and shade, while performing intense physical labor. Additionally, tight project deadlines may compel them to work during peak heat hours, further exacerbating heat-related productivity loss. These findings highlight the heightened vulnerability of construction workers to heat exposure and underscore the need for effective heat protection measures.

Variations in productivity loss among construction worker subgroups

Impact of age on productivity loss

Age plays a significant role in productivity loss under heat exposure. Among workers aged 38 and older in this review, 61% reported productivity loss, likely due to age-related declines in cardiovascular function, reduced thermoregulation, and lower resilience to heat stress. Older workers were more vulnerable to extremely high-temperature conditions, probably due to the following reasons. First, the function of the cardiovascular system gradually declines with age. The heart’s ability to pump blood is weakened, and blood vessel elasticity is reduced. These changes make it difficult for the body to regulate temperature effectively in high temperatures. As a result, older workers are more prone to problems such as poor blood circulation and increased heart rate at high temperatures, leading to a decline in physical strength and cardiovascular aging [32]. Second, older workers are less resilient to heat exposure. Their sweat glands function less efficiently than younger people so they may lose heat less in high temperatures. In addition, older workers have decreased kidney function, which affects the body’s water balance and electrolyte regulation. In high temperatures, they are more prone to dehydration and electrolyte imbalances, leading to fatigue, dizziness, and decreased concentration, affecting work productivity [33]. These results highlight the importance of targeted interventions for older workers, including personalized cooling strategies and work-rest cycles to mitigate productivity loss.

Gender differences in heat-related productivity loss

Our analysis revealed gender differences in heat-induced productivity loss. Among mixed-gender groups, 74% experienced productivity loss, compared to 45% in male-only groups. This suggests that women may be more sensitive to heat exposure, possibly due to greater cardiovascular strain and orthostatic intolerance. From a physiological point of view, women are more prone to cardiovascular strain and orthostatic intolerance when working at high temperatures. Women are five times more likely to have symptoms related orthostatic intolerance than men of similar age and health status. Symptoms of orthostatic intolerance, including dizziness, lightness of the head, fatigue, headache, and upright syncope, can affect a woman’s ability to work and lead to reduced productivity [34]. The number of women in the construction industry is relatively small compared to the number of men. Most female construction workers are unskilled and illiterate, and typically receive lower wages. Their primary roles involve assisting male workers by carrying heavy loads such as cement bags, bricks, and other materials. Research indicates that work-related stress, sexual harassment, and gender-based discrimination among female construction workers negatively affect their mental and physical health, leading to reduced work capacity and less output [35].

Geographic differences in heat-related productivity loss

Surprisingly, productivity losses were lower in middle-low-income countries (58%) compared to high-income countries (64%). This contradicts previous research, which suggests that less affluent regions are more vulnerable to heat exposure. Studies have shown that global warming often affects less affluent areas more severely, as they may lack essential facilities and resources like high-tech tools, a skilled workforce, or financial means to adapt to such challenges [36]. However, our review presents a different perspective. Several factors may explain this discrepancy. First, construction sites in high-income countries often adhere to stricter safety and health regulations, requiring workers to wear heavy protective gear. While this clothing provides necessary safety, it can impair the body’s ability to regulate temperature in high-heat environments, increasing workers’ risk of heat exposure and reducing productivity [37]. Second, high-income countries typically enforce stringent labor protection regulations, including temperature thresholds that mandate work stoppages during extreme weather. While these regulations protect workers from heat-related health risks, they also reduce the number of working hours, directly impacting overall productivity [38]. In addition, only a small number of high-income countries were represented in this review—specifically, three articles from Australia, Sweden, and Italy—limiting the ability to fully reflect the characteristics of all high-income countries. Future studies should include a broader range of countries and regions to better understand the global impact of heat exposure on construction productivity.

Temporal trends in productivity loss due to heat exposure

We divided the data into two groups based on the research period: one covering studies conducted before 2016 and the other after 2016. The results showed that 63% of construction workers experienced productivity losses before 2016, compared to 57% after 2016, indicating a decline in productivity loss since 2016. This decrease can be attributed to the digitalization of the global construction industry, which has shifted from traditional construction methods to those based on digital technologies. Since 2016, the pace of digital transformation in the construction industry has accelerated, with wearable technology and health monitoring devices becoming widely adopted. In 2018, digital technologies such as construction site robots and lean construction technology emerged [39, 40]. In 2019, blockchain technology and BIM5D gradually emerged in the construction industry, followed by the application of 3D printing [41, 42] and the Internet of Things to construction projects in 2020 [43, 44]. These digital technologies have been increasingly integrated into the construction phase, positively affecting productivity and helping to mitigate the productivity loss caused by high temperatures for construction workers [45].

Influence of temperature thresholds on productivity

Different levels of temperature exposure led to varying rates of productivity loss among construction workers. When the average WBGT exceeded 28°C and the ambient temperature surpassed 35°C, 62% of workers experienced productivity losses, a higher percentage than when the WBGT was below 28°C and the ambient temperature below 35°C. This demonstrates that high temperatures have a significant impact on labor productivity, with higher temperatures leading to greater losses. Zhang et al. found that the “ideal temperature” for peak productivity among construction workers is 24.90℃, with optimal performance occurring around 25.0℃ [46]. In high-temperature conditions, the physical demands of strenuous construction work can lead to mental fatigue or anxiety, reducing focus and lowering productivity [47]. As temperatures continue to rise, individual work capacity may decrease, with a broader negative impact on construction projects, further exacerbating productivity losses for construction companies [48].

Comparison of self-reported and measured productivity loss

There was a discrepancy between self-reported and actual assessments of productivity losses among construction workers. The self-reported figure was 67%, slightly higher than the overall combined effect size, while the actual measured result was 13%, lower than the overall combined effect value. The higher self-reported result may be attributed to subjective perception and recall bias, as workers’ reports were based on their personal experience, which can introduce certain inaccuracies [49]. In addition, social expectations may influence workers’ responses, making them more likely to report productivity losses to highlight challenging working conditions or advocate for improved measures [50]. On the other hand, actual measurements provide more objective data, capturing only significant productivity losses, which may overlook minor declines. Most of the studies in this review relied on self-reported data due to its convenience, affordability, and widespread public acceptance [51]. In contrast, actual measurements require specialized technical equipment and rigorous experimental designs, which may be difficult to implement in some studies [52]. Consequently, the number of studies using actual measurements was limited. Future studies could benefit from combining self-reported data with field measurements to generate more comprehensive data and better validate the findings.

Limitations

First, many existing studies were concentrated in a few countries or regions, such as Australia, Sweden, and Italy, making it difficult for the findings to be globally representative. Climatic conditions, working environments, and labor protection regulations vary significantly across regions, so studies based on limited geographic samples may not provide a comprehensive view of the challenges faced by all construction workers. Second, the effects of heat exposure on work productivity may vary over time, with short-term and long-term exposure having potentially different impacts. However, many studies did not distinguish between these temporal effects in detail, which limits our understanding of the full spectrum of heat exposure impacts. In addition, individual differences are also important to consider. Factors such as age, health status, work experience, work intensity, and personal protective measures can affect how workers respond to heat exposure [53, 54]. However, most studies have not adequately accounted for these individual differences. In this review, only gender and age were considered, which may have caused the results to not fully reflect the diversity of construction workers. Lastly, cross-sectional studies without comparison groups hindered the establishment of a definitive causal relationship between heat exposure and productivity loss. In summary, although this review offers valuable insights into the effects of heat exposure on construction workers’ productivity, the results should be interpreted with caution. Future studies should address these limitations to improve the comprehensiveness and accuracy of the findings.

Conclusions

The meta-analysis results from various regions indicated that high temperatures significantly affect the productivity of construction workers, with 60% experiencing a loss of work productivity under high-temperature conditions. As maximum temperatures continue to rise in the context of climate change, the likelihood of further declines in outdoor productivity during hot weather remains high. The study also found that female workers aged 38 years and older were more sensitive to heat exposure, highlighting the need for business managers to address inequality by giving special attention to this vulnerable group.

Information technology can help reduce productivity loss to some extent, so it is essential to actively explore and develop innovative technologies and equipment designed for construction work in hot weather. The use of intelligent wearable devices, such as vests with built-in cooling systems and wearable sensors, should be expanded. In addition, in middle-low-income countries, efforts should be made to raise awareness of heat exposure and its effects, as well as to improve labor insurance policies and legal standards for working in high-temperature environments. Additionally, work and social conditions need to be improved to better protect workers. This includes ensuring constant access to drinking water, allowing workers to adjust their work pace based on their conditions, providing shaded areas, and maintaining adequate sanitation facilities. Improving the psychosocial work environment is also important. This can be achieved by optimizing working hours, increasing worker autonomy and control, providing social support, and ensuring a balance between effort and reward.

Abbreviations

CI

Confidence Interval

WBGT

Wet Bulb Globe Temperature

PRISMA

Preferred reporting items for Systematic Reviews and Meta-analyses

PROSPERO

Prospective Register of Systematic Reviews

JBI-MAStARI

Joanna Briggs Institute Meta-Analysis of Statistics Assessment and Review Instrument

AACE

American Association of Cost Engineers

Hothaps

High Occupational Temperature Health and Productivity Suppression

ROB

Risk of Bias

GDP

Gross Domestic Product

Author contributions

S.H., and L.D. designed and implemented the study, analyzed the data, interpreted the results, and drafted the manuscript. Y.W. and J.X. aided in the study procedures and assisted in manuscript preparation. All authors read and approved the final manuscript.

Funding

No funds will be received for this review.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethical approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.